A technology of identifying music by listening to a song is a technology in which after collecting a recording, a local client extracts an audio fingerprint of the recording, uploads the audio fingerprint to a server, then matches the audio fingerprint with data of the server by means of audio fingerprint matching, and returns a query result. The audio fingerprint can represent features, including tone, loudness, timbre and the like, of a song, and the information may be obtained by means of reporting of a service client of music playback application software.
Currently, the solution generally used to implement music playback by using the technology of identifying music by listening to a song is as follows: when a user listens to a song and records an audio file, related audio information is collected by using a service client program, the related audio information is reported to a data analysis server, and after obtaining the audio information by means of matching, the data analysis server provides downloading and playing services to the user.